Claude Fable 5.2 (max): benchmark scores, price and ranking

Claude Fable 5.2 (max) from Anthropic ranks #1 of 827 LLMs on the LLMs Tiger aggregate benchmark score (63.4), based on 16 benchmark results captured up to 2026-11-18.

Organisation
Anthropic
Type
Frontier model
Open weights
No
Released
2026-11-09 (source: epoch)
Cheapest API price
$8.00 in / $40.00 out per million tokens (via anthropic, observed 2026-11-18T15:47Z)
Aggregate rank
#1 of 827 (score 63.4)

Relative to other models, Claude Fable 5.2 (max) is strongest on SciCode (#1 of 113) and weakest on PostTrainBench (#2 of 13).

Across the 16 benchmarks where at least ten models are scored, Claude Fable 5.2 (max) is in the top 10% on 15 and in the top half on 16.

Its cheapest listed input price, $8.00 per million tokens, is lower than that of 6% of the 443 models with a listed price.

Claude Fable 5.2 (max) benchmark results

Claude Fable 5.2 (max) benchmark scores, ranked against all models measured on each benchmark
BenchmarkScoreRankSourceCaptured
SciCode67.7%#1 of 113Epoch AI Benchmarking Hub2026-11-18
Chess Puzzles79.0%#1 of 84Epoch AI Benchmarking Hub2026-11-18
SimpleQA Verified94.0%#1 of 83Epoch AI Benchmarking Hub2026-11-18
FrontierMath T1-395.1%#1 of 78Epoch AI Benchmarking Hub2026-11-18
CritPt35.4%#2 of 138Epoch AI Benchmarking Hub2026-11-18
FrontierMath T4100.0%#1 of 60Epoch AI Benchmarking Hub2026-11-18
Furniture Assembly90.0%#1 of 36Epoch AI Benchmarking Hub2026-11-18
Mystery Games90.0%#2 of 69Epoch AI Benchmarking Hub2026-11-18
ARC-AGI-292.5%#4 of 97ARC Prize official leaderboard2026-11-18
FrontierSWE69.9%#1 of 24Epoch AI Benchmarking Hub2026-11-18
ProofBench100.0%#2 of 47Epoch AI Benchmarking Hub2026-11-18
GPQA Diamond93.9%#15 of 302Epoch AI Benchmarking Hub2026-11-18
OTIS Mock AIME99.7%#15 of 206Epoch AI Benchmarking Hub2026-11-18
EBR-Bench90.5%#2 of 27Epoch AI Benchmarking Hub2026-11-18
CursorBench60.0%#2 of 22Epoch AI Benchmarking Hub2026-11-18
PostTrainBench44.6%#2 of 13Epoch AI Benchmarking Hub2026-11-18

About Claude Fable 5.2 (max)

Claude Fable 5.2 (max) is a frontier model from Anthropic. It was released on 2026-11-09, according to Epoch AI. The cheapest API listing we track charges $8.00 per million input tokens and $40.00 per million output tokens, through anthropic. We hold 16 benchmark results for it across 3 areas, and 15 are in the top tenth of their boards.

Where Claude Fable 5.2 (max) stands

On our aggregate score, Claude Fable 5.2 (max) is #1 of 827 ranked models, in the top ten. Its score is 63.4; the scale is centred on 50, so it sits 13.4 points above the typical model. That beats 100% of the ranked field. Among the 73 ranked Anthropic models it is #1. The next one down is Claude Opus 5.5 (max) at 63.1.

Claude Fable 5.2 (max), benchmark by benchmark

SciCode is where Claude Fable 5.2 (max) does best: 67.7%, first of 113. The next result is Claude Opus 5.5 (max) at 66.9%, 0.8 points behind. It covers scientific research code. The board median is 46.1%, so Claude Fable 5.2 (max) is 21.6 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

Chess Puzzles is where Claude Fable 5.2 (max) does best: 79.0%, first of 84. The next result is Gemini 4 Carbon (high) at 78.0%, 1.0 points behind. It covers chess tactics puzzles. The board median is 33.0%, so Claude Fable 5.2 (max) is 46.0 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

SimpleQA Verified is where Claude Fable 5.2 (max) does best: 94.0%, first of 83. The next result is GPT-6.2 Astra (max) at 78.5%, 15.5 points behind. It covers short factual questions. The board median is 41.4%, so Claude Fable 5.2 (max) is 52.6 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

FrontierMath T1-3 is where Claude Fable 5.2 (max) does best: 95.1%, first of 78. The next result is GPT-6 Astra (max) at 93.7%, 1.4 points behind. It covers research-level maths, tiers 1-3. The board median is 61.6%, so Claude Fable 5.2 (max) is 33.5 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

On CritPt, Claude Fable 5.2 (max) scores 35.4% and ranks #2 of 138. Only one model scores higher, led by Gemini 4 Carbon (high) at 36.0%, which is 0.6 points ahead. It covers unpublished research-level physics problems. The board median is 2.0%, so Claude Fable 5.2 (max) is 33.4 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

FrontierMath T4 is where Claude Fable 5.2 (max) does best: 100.0%, first of 60. The next result is GPT-6.1 Sol (max) at 100.0%, 0.0 points behind. It covers research-level maths, tier 4. The board median is 32.9%, so Claude Fable 5.2 (max) is 67.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

On OTIS Mock AIME, Claude Fable 5.2 (max) scores 99.7% and ranks #15 of 206. Only 14 models score higher, led by Claude Fable 5.1 (max) at 100.0%, which is 0.3 points ahead. It covers competition maths (AIME 2024-25 style). The board median is 66.4%, so Claude Fable 5.2 (max) is 33.3 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

On EBR-Bench, Claude Fable 5.2 (max) scores 90.5% and ranks #2 of 27. Only one model scores higher, led by GPT-6.2 Astra (max) at 92.4%, which is 1.9 points ahead. It covers evidence-based reasoning. The board median is 39.5%, so Claude Fable 5.2 (max) is 51.0 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

On CursorBench, Claude Fable 5.2 (max) scores 60.0% and ranks #2 of 22. Only one model scores higher, led by Gemini 4 Carbon (high) at 60.4%, which is 0.4 points ahead. It covers real-world coding-agent tasks. The board median is 46.5%, so Claude Fable 5.2 (max) is 13.5 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

PostTrainBench puts Claude Fable 5.2 (max) at 44.6%, #2 of 13. That beats 85% of the models tested, but 0.0 points short of Claude Fable 5.1 (max) (44.6%). The results either side are Claude Fable 5.1 (max) (44.6%) and Claude Opus 5.5 (max) (43.8%). It covers post-training a smaller model. The board median is 38.8%, so Claude Fable 5.2 (max) is 5.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-11-18.

The full table above also lists the other 6 results, including boards with fewer than ten models, where a rank says little.

Strengths and weaknesses

Software engineering: strong, ahead of 95% of the field on average across 3 benchmarks. Its best board here is SciCode (#1 of 113) and its weakest is CursorBench (#2 of 22). Reasoning, maths and science: strong, ahead of 96% of the field on average across 7 benchmarks. Its best board here is FrontierMath T1-3 (#1 of 78) and its weakest is OTIS Mock AIME (#15 of 206). General knowledge and chat: strong, ahead of 95% of the field on average across 6 benchmarks. Its best board here is Chess Puzzles (#1 of 84) and its weakest is PostTrainBench (#2 of 13).

Price and value

The median input price across the 443 priced models is $0.482 per million tokens, so Claude Fable 5.2 (max) costs more than the typical listing. Models that score within 1.5 points or better and cost less: Gemini 4 Carbon (high) ($2.00, 62.4), Claude Opus 5.5 (max) ($4.00, 63.1). Similar scores cost more on GPT-6 Astra (max) ($10.00).

Release history

The Anthropic model released before it was Claude Haiku 5.5 (max) on 2026-10-07, which scores 54.3. The next one was Claude Fable 5.2 (with fallback, max) on 2026-11-09, at 52.4.

Reading these numbers

The 16 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (15), ARC Prize official leaderboard (1). 1 of them come from an official leaderboard and 15 from datasets or third-party runs. The leaderboard also lists 6 other reasoning-effort settings of this model, which redirect here. This page shows the highest.

Claude Fable 5.2 (max) is also listed at 6 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.

Models ranked nearby

More from Anthropic

Compare it on the live leaderboard → · How the aggregate score works